We conducted an epidemiologic case-control study of thyroid cancer in women aged 40 and under to test the hypothesis that endogenous hormones may relate to the development of this disease, since the only known cause of thyroid cancer, ionizing radiation, does not account for the striking female over male excess. When compared to neighbour controls women with thyroid cancer more often had a history of benign hyperplastic thyroid disease (Relative Risk (RR) = 14.5; P less than 0.01) and more often had ever been pregnant (RR = 2.1; P = 0.04). Both these findings were consistent with findings of previous studies. After eliminating women with a history of hyperplastic thyroid disease from the analysis we found a strong association with miscarriage as the outcome of the first pregnancy (RR = 11.5; P less than 0.01), and we suspect that this factor may be another indicator of thyroid abnormality. An independent and increasing risk was observed with an increase in the total number of pregnancies after excluding women with prior thyroid disease and those whose first pregnancy ended in a miscarriage. The RR for 4 or more pregnancies was 6.3 (P = 0.03). Prior exposure to radiation therapy was not an important factor in our study of young women; this suggests that the emphasis in future studies of thyroid cancer must shift to study other types of risk factors.